The Enterprise AI Earnings Test: What Salesforce, Marvell, and Snowflake Must Prove Today
There is a phrase that gets used a lot in technology investing right now: "AI monetization." The idea is simple enough. Hundreds of billions of dollars have been poured into AI infrastructure - chips, data centers, power, cooling - and the question that has been building for two years is whether the software layer sitting on top of all that hardware can actually convert the investment into recurring revenue. Today, May 27, 2026, three companies are reporting earnings after the market close that will go a long way toward answering that question. Salesforce, Marvell Technology, and Snowflake are not the same kind of business, but they share a common thread: each one has staked its near-term growth story on AI, and each one is being asked to prove it in the numbers.
Salesforce: The Agentforce Moment
Salesforce is the most watched of the three. The company reports its first quarter of fiscal 2027 results this afternoon, with analysts expecting revenue of $11.05 billion and earnings per share of $3.12. Those are not the numbers that matter most. What matters is Agentforce, the company's autonomous AI agent platform that CEO Marc Benioff has been positioning as the defining product of the next decade of enterprise software.
The setup heading into today is complicated. In the fourth quarter of fiscal 2026, Agentforce annual recurring revenue reached $800 million, up 169% year over year, with 29,000 deals closed - up 50% quarter over quarter. Those are genuinely impressive numbers for a product that barely existed 18 months ago. But Bank of America reinstated Salesforce with an Underperform rating and a $160 price target last week, calling it a "structural reset story" and arguing that the transition to agentic AI could actually compress near-term revenue as customers shift away from traditional seat-based licensing. The stock closed at $179.51 on Monday, down from highs above $300 in 2023.
The bear case is not irrational. If AI agents can do the work of multiple human users, the seat-based model that has underpinned Salesforce's revenue for two decades faces structural pressure. Benioff's counter-argument is that agents expand the addressable market rather than cannibalize it - that companies will deploy agents to do work that was never being done by software at all. Today's results will not resolve that debate definitively, but the trajectory of Agentforce ARR and the Q2 guidance will tell investors whether the ramp is accelerating or plateauing.
Marvell: The Chip That Outran Its Analysts
Marvell Technology has done something unusual in 2026: it has outrun Wall Street. The stock has more than doubled year to date, reaching $196 per share, while the consensus analyst price target sits at roughly $150 - nearly 24% below where shares are trading. That gap is not a sign of irrational exuberance. It is a sign that the revision cycle has been moving faster than analysts can update their models.
The catalyst has been real. In March, Nvidia announced a $2 billion investment in Marvell and integrated the company into NVLink Fusion, its rack-scale AI infrastructure platform. In April, reports emerged that Google is in discussions with Marvell to co-develop custom AI chips. CEO Matt Murphy has raised his fiscal 2027 revenue outlook three times in six months, from $9.5 billion to $10 billion to approaching $11 billion, each revision driven by accelerating bookings from hyperscalers. He guided Q1 fiscal 2027 revenue of $2.4 billion at the midpoint, implying 27% year-over-year growth, with non-GAAP EPS of $0.79.
The question today is not whether Marvell is executing. It clearly is. The question is whether Murphy signals a fourth consecutive upward revision, or whether the cadence of upgrades has peaked. His commitment to Q4 fiscal 2027 revenue exceeding $3 billion is the benchmark. If Q2 guidance implies a trajectory consistent with that exit rate, the stock holds its premium multiple. If it implies deceleration, the gap between the $196 price and the $150 Street consensus closes to the downside.
Snowflake: The Data Platform in the Middle of Everything
Snowflake occupies a different position in the AI stack. It is not a chip company and it is not a CRM platform. It is the place where enterprise data lives, and the thesis is that AI makes that data more valuable, not less. In February, the company guided fiscal 2027 product revenue of $5.66 billion, above analyst estimates, and disclosed that it had signed the largest deal in its history - over $400 million - without naming the client. It also struck separate $200 million multi-year deals with both OpenAI and Anthropic to integrate their models into its platform.
The risk for Snowflake is the same one that has haunted enterprise software broadly: the fear that AI tools reduce the need for the underlying data infrastructure rather than expanding it. D.A. Davidson analyst Gil Luria pushed back on that narrative after the Q4 results, arguing that "as Snowflake continues to accelerate throughout the rest of the year, investors will realize Snowflake is benefiting significantly from the growth of AI." Today's Q1 results, with analysts expecting EPS of $0.14, will test that thesis against actual consumption data.
What the Market Is Really Watching
The broader significance of today's triple earnings report is what it says about the next phase of the AI trade. For the past two years, the market has rewarded the infrastructure layer - Nvidia, Micron, Marvell - with extraordinary valuations because the demand for compute and memory has been unambiguous. The software layer has been more contested. Enterprise software stocks have underperformed the AI infrastructure names because the monetization story has been harder to verify in real-time revenue data.
Today is one of the first days where that verification happens at scale. Salesforce, Marvell, and Snowflake together represent the compute, the data, and the application layers of the enterprise AI stack. If all three report results that confirm the AI monetization thesis - accelerating ARR, strong guidance, evidence that enterprise customers are committing real budget to AI tools - the rotation from infrastructure to software that many investors have been anticipating could finally have a catalyst. If the results are mixed or disappointing, the debate continues. Either way, the market will have more information tonight than it did this morning, and that is exactly what earnings season is supposed to provide.
